1. Technical Field
The present invention relates to an image processing system, an image processing method, and a computer-readable medium.
2. Related Art
A method for resolution conversion of an image has been known (e.g., see Patent Document 1). The method includes reading high-resolution image-encoding data, entropy decoding the data, inverse quantizing the decoded data, performing a matrix operation of the inverse-quantized data and a resolution-conversion matrix that is generated from a discrete-cosine-transform matrix according to a desired resultant resolution, in order to obtain a low resolution discrete-cosine-transformed data. The obtained data is further quantized and the quantized data is entropy encoded.
Another known image processing apparatus divides an original image data into a multiple blocks that includes a plurality of pixels and performs image processing on frequency data that is obtained by converting the blocks of the image data into spatial-frequency components (e.g., see Patent Document 2). In the image processing apparatus, each piece of frequency data in each block is divided into regions according to the frequency level. Moreover, in the image processing apparatus, frequency data of each region is summed starting from the region with a low frequency until the sum reaches or surpasses a threshold value, and a processed region is set to include the regions situated before the region at which the sum reaches or surpasses the threshold value. Data is then selected from the processing region, and an operation is performed on the selected data.
As for the above mentioned examples of the related art, Patent Document 1 is Japanese Patent Application Publication No. H5-316357 and Patent Document 2 is Japanese Patent No. 3722169.
However, according to the techniques described in Patent Document 1 and Patent Document 2, all the pixel data in a whole image area are to be operated. Therefore, a large amount of operations has to be performed and it takes a long time to complete a compression processing.